Biomass allocation patterns and allometric relationships between components of the androdioecious Acer tegmentosum

Abstract

Key message

We present comparisons about biomass allocation between males and hermaphrodites of androdioecious Acer tegmentosum Maxim.. Different biomass allocation patterns were found, and males were shown to have a larger investment into coarse roots and foliage.

Context

Sexual dimorphism in differences of reproductive costs between genders has been widely reported for trees, but we still know little about allometric relationships between tree components in both genders.

Aims

We present biomass allocation patterns and relationships between components of the androdioecious A. tegmentosum in a broad-leaved mixed forest in northeastern China. The objectives of this study were to examine how gender affects the biomass structure of androdioecious species and how the gender-related reproductive efforts affect the allometric relationships.

Methods

We harvested 31 hermaphrodite and 29 male A. tegmentosum trees and opted for diameter at breast height, tree height, and crown length as the independent variables and various biomass components as the dependent variables. Five types of function were used to model allometry equations.

Results

Biomass allocation between genders was different, and the best biomass model for each biomass component varies between genders. Males have a higher investment in foliage and coarse root biomass than hermaphrodites, and hermaphrodites invested more in reproduction than males.

Conclusion

Biomass equations are strongly gender-related. Males tended to invest a larger fraction of the vegetative biomass into leaves and coarse roots.

1 Introduction

In a sexual dimorphic reproductive system, androdioecy is very rare. This system consists of only two types of individual plants in breeding populations, i.e., male and cosexual plants (Yampolsky and Yampolsky 1922). Male plants produce male flowers, while cosexual plants produce perfect flowers. We chose a “hermaphrodite” to represent the second kind of plant. Lloyd (1979)) pointed out that androdioecious plants have two forms in nature: (1) true functional androdioecy, where perfect flowers carry both female and male functions, and (2) morphological androdioecy, where the anthers of perfect flowers are indehiscent or barren, i.e., they are functionally dioecious or monoecious (Lloyd 1979; Anderson and Symon 1989; Swensen et al. 1998). Only seven species are known to exhibit true functional androdioecy (Vassiliadis et al. 2000). No matter which kind of androdioecy, hermaphrodites have to allocate their limited resources to both flowering and setting fruits, while males produce flowers only.

The theory of plant life history is largely concerned with optimal resource allocation between male and female functions (Allen and Antos 1993; Rocheleau and Houle 2001). It is assumed that there is a resource tradeoff between reproductive and vegetative distribution (Popp and Reinartz 1988; Allen and Antos 1993; Antos and Allen 1999; Obeso 2002; Torimaru and Tomaru 2012). A number of comparative studies have examined sexual dimorphism, although some disagreement persists with this issue. Most reports indicate that a higher cost of reproduction should have some consequences, such as lower vegetative growth; hence, males will have a higher growth rate than females (Obeso et al. 1998; Antos and Allen 1999; Obeso 2002; Cornelissen and Stiling 2005). But the cost of fruit production seems to be not too high to limit vegetative growth in other species (Sakai 1990). Female plants have a powerful feedback mechanism to compensate for the higher reproductive effort by saving resources for stem and height growth (Petzold et al. 2013). For example, Nicotra (1999) reported that females grow faster than males. Sexual dimorphism may contribute to some intersexual secondary differences, which can be explained by the higher reproductive costs incurred by hermaphrodites than by males.

Most investigations have focused on reproductive costs, but little attention has been paid to sexual differences in biomass allocation patterns and relationships regarding the main tree architecture of cosexual species. Individuals may vary in their relative allocation to male and female functions (Lloyd 1979). Perhaps one of the most frequently observed differences between genders is that females allocate a greater proportion of biomass to reproduction components than do males (Meagher and Antonovics 1982; Popp and Reinartz 1988). Zhang et al. (2012) studied the biomass allocation of two diecious shrub species of the genus Rhamnus. They found that females of these two species had more foliage, branch, and trunk biomass than males. Michiko et al. (2015) analyzed the allometric relationship between tree size and vegetative components of both genders. Their results demonstrate that females have greater leaf biomass than males. There are still only a few studies revealing biomass allocation patterns and allometric relationship of androdioecious species.

In this study, we examined a basic biomass allocation pattern of the androdioecious Acer tegmentosum Maxim., relating biomass components for both genders. The specific objectives of this study were to investigate (1) whether the biomass structure of androdioecious species is gender-related and (2) whether the differences in reproductive efforts affect the allometric relationships of both genders.

2 Methods

2.1 Study site

The study was conducted at the Jiaohe experimental forest in Jilin province (43° 57.5′ N, 127° 44.1′–127° 44.7′ E) at an average elevation of 459 m, with a mean annual temperature of 3.8 °C. The region has a monsoon climate; precipitation is low in winter but at higher elevations in summer it is very high, with a mean annual precipitation as high as 695.9 mm. The soil is a dark brown forest soil. The study site is a multi-species near-mature forest dominated by Pinus koraiensis Sieb., Fraxinus mandshurica Rupr., Tilia amurendsis Rupr., and Acer mono Maxim..

The androdioecious species A. tegmentosum in this study is common as a small tree at the study site. It is widely distributed in the conifer and broad-leaved mixed forests in northeastern China. Both genders of A. tegmentosum blossom in early spring, with racemes of green flowers. For hermaphrodites, the fruits come with dipterous appearance and change from green to yellowish-brown when ripe in late August.

2.2 Sampling and measurements of biomass components

In May 2014, after identifying the sex of 60 flowering A. tegmentosum trees (29 males, 31 hermaphrodites) and marking their location, we counted the number of flowers of each tree. We took about 50 flowers to the laboratory, oven-dried them at 80 °C for 48 h, weighed their dry mass, and then estimated the total flower biomass.

All the sample trees (29 males and 31 hermaphrodites) were cut at ground level after measuring the diameter at breast height and diameter at tree base. After cutting, we dissected the entire plant into different components, i.e., vegetative parts such as foliage, branches, stem wood, and coarse roots, and the branches were further classified as level 1, 2, and 3 branches, level 3 being the smallest and level 1 the largest (Fig. 1), as well as fruits for hermaphrodites. We measured the fresh weights for each component separately in the field to the nearest 0.1 kg. We measured only coarse roots (diameter ≥ 5 mm), because harvesting fine roots is practically impossible and they have little impact on total root biomass. The roots were cleaned and their fresh mass weighed. Simultaneously, 0.5–1 kg representative samples of each component were taken to the laboratory, oven-dried to a constant mass, and weighed for determination of their moisture content. Total dry matter of each component was calculated by the ratio of the samples (dry matter to fresh weight). Statistical summaries of sampled trees are presented in Table 1.

Fig. 1

Simplified scheme of the entire tree. We dissected each sample tree into different components including foliage, branches, stem wood, and coarse roots, and the branches were further classified as level 1, 2, and 3 branches, level 3 being the smallest and level 1 the biggest

Table 1 Statistical summary of sample tree variables and biomass data for different genders

2.3 Data analysis

Analysis of covariance (ANCOVA) with DBH and gender as covariates was used to examine sexual differences in biomass proportion. The observations meet the assumptions of independence with equal error variances. A general non-linear biomass equation ($Y=p{X}_{{}_1}^{p_1}{X}_2^{p_2}{X}_3^{p_3}\cdots {X}_n^{p_n}\theta$) was used to develop equations for each gender relating tree component and belowground biomass (Xiao and Ceulemans 2004; Mesele et al. 2013; Nazanin and Taraneh 2013; Rosta et al. 2014; Fabiola et al. 2015), where Y is dry matter of a component and X _i the independent variables, p ₀ − p _n are model coefficients, and θ represents the multiplicative error term. For each biomass component, equations were developed with different combination of independent variable: (i) DBH, (ii) tree volume index (DBH²H dm³), and (iii) the combination of DBH, H, and CL, where DBH is the diameter at breast height, H the tree height, and CL the crown length. After modelling, we checked whether the residuals were normally distributed.

Model selection was guided by an information criteria approach (Burnham and Anderson 2002). Employing an assumption of multivariate normality, Akaike’s information criterion (AIC) is a way of selecting an equation from a set of alternative equations and was used to decide on the best model (Akaike 1974). Root mean square error (RMSE) was also used for the selection of the best equation (Fayolle et al. 2013; Vahedi 2016). The formulas are defined as:

$$ \mathrm{A}\mathrm{I}\mathrm{C}=2k+n \ln \left(\frac{1}{n}{\displaystyle \sum_{i=1}^n{\left({y}_i-{\widehat{y}}_i\right)}^2}\right) $$

(1)

$$ \mathrm{RMSE}=\sqrt{\frac{1}{n}{\displaystyle \sum_{i=1}^n{\left({y}_i-{\widehat{y}}_i\right)}^2}} $$

(2)

where k is the number of parameters, n the number of trees, y _i the observed value, and ŷ _i the estimated value for a tree i, respectively.

If there was no difference in the model form between genders, we demonstrated the occurrence of different allometric patterns by randomly selecting 1000 numbers from the minimum to the maximum independent variables we sampled, and then we calculated the values of biomass for both genders with the selected best equations. A general t test was used to check the differences between the populations of the two biomass groups. After all the data were transformed by natural logarithms, we tested the statistical differences between regressions for component biomass with the Chow test (Dantas et al. 2013; Khanam et al. 2015; Jadin et al. 2016).

All analyses were performed with R, version 3.2.3 (R Development Core Team 2014). R package nls was used for modelling, and the R library “gap” was used for Chow test analyses.

3 Results

3.1 Vegetative biomass components in both genders as a proportion of total biomass

No significant difference was found in total biomass between male and hermaphrodites (F = 3.78, p = 0.571). We examined the component biomasses of both hermaphrodites and males as a proportion of total biomass and found it gender-specific. Compared with males, hermaphrodites have a smaller ratio of foliage biomass to total biomass, while in terms of belowground biomass, the percentage of coarse roots biomass is higher in males than in hermaphrodites (Fig. 2).

Fig. 2

Comparison of the biomass of selected components between hermaphrodite and male trees. Note 1: Proportion of biomass is defined as the ratio of biomass from corresponding components to total biomass. Note 2: The bars are standard deviations and different letters over the columns indicate significant differences (p < 0.001)

3.2 Gender-related biomass equations

All allometric equations of both genders, based on DBH, are significant (Tables 2 and 3). According to the evaluation of AIC and RMSE, the forms of the best model for stem wood and total biomass equations are different between genders, and the DBH²H independent variable model was found to be the best fit for foliage, branch, and aboveground and belowground biomass for both genders we sampled. In terms of flower and branch biomass, the DBH model proved to be best. Both the Chow test and t test showed significant difference between genders for the same equation form, except for branch biomass (Table 4). These results indicated that the biomass models were significantly different between male and hermaphrodite trees, suggesting that biomass equations are gender-related for flower, foliage, branch, and aboveground and belowground components.

Table 2 Regression coefficients (p ₀ − p ₃) for different equations used to estimate component biomass (g) with DBH (cm), DBH²H (dm³), H (m), and CL (m) for male Acer tegmentosum trees

Table 3 Regression coefficients (p ₀ − p ₃) for different equations used to estimate component biomass (g) with DBH (cm), DBH²H (dm³), H (m), and CL (m) for hermaphrodite Acer tegmentosum trees

Table 4 Summary of the allometric patterns of component biomass in males and hermaphrodites using the Chow test and t test

3.3 Allometry in vegetative/reproductive components

We analyzed the relationships of vegetative components and independent variable for both sexes (see Fig. 3). The results showed that the biomass of the various parts differs in their allometric relationships. For vegetative components, the values of both parameters (multiplier and power parameter) are higher for males than for hermaphrodites in foliage and coarse root biomass (Fig. 3 and Table 4). In terms of aboveground biomass, males have a lower power parameter (0.890) but higher multiplier (47.5). There were no significant differences in branch biomass. Due to their higher investment in foliage and coarse root biomass, male trees allocate more resources to vegetative biomass.

Fig. 3

Vegetative components as a function of DBH or DBH²H for both male and hermaphrodite trees. For each regression of the components, the bivariate lines were fitted on the basis of the best model selected in Tables 2 and 3

We defined reproductive biomass as the flower biomass for males and the sum of flower and fruit biomass for hermaphrodites. We found no significant differences in the number of flowers in both sexes. For total reproductive biomass, hermaphrodites invested more in reproduction than males in all the DBH classes observed (Fig. 4).

Fig. 4

Reproductive components as a function of DBH or DBH²H for both male and hermaphrodite trees. The reproductive biomass was defined as the biomass of flowers for males while the sum of flower and fruit biomass for hermaphrodites. The bivariate lines were fitted for each regression

4 Discussion

4.1 Gender-dependent biomass allocation pattern

Many studies reported sexual dimorphism, which is a phenomenon frequently associated with the differences in gender allocation patterns, especially in dioecious species (Sakai 1990; Allen and Antos 1993; Nanami et al. 2005; Zhang et al. 2014). Obeso (1998) reported that females allocate more resources to reproduction than males and exhibit a slower growth rate (Obeso et al. 1998). Shea et al. (1993) studied water tupelo trees (Nyssa aquatic Linnaeus.) and found that males were significantly larger in height and basal area than females, suggesting that males allocate more resources to vegetative growth than females. Similarly, Zhang et al. (2014) reported higher rates of stem grow in males for the diecious species F. mandshurica Rupr..

Reproductive costs of females are usually higher than those of males, which may contribute to differences in resource allocation. Females usually put more resources to defense and reproduction, while males allocate more resources towards vegetative growth (Verónica and Rodolfo 2010; Petzold et al. 2013). Life history theory of plants depends heavily upon how they allocate their limited resources to vegetative growth, reproduction, defense, and other capabilities (Roff 1993; Bazzaz et al. 2000). Therefore, there might be a tradeoff between reproductive and vegetative resource allocation, where vegetative growth is affected by reproductive costs (Popp and Reinartz 1988; Rocheleau and Houle 2001). We assume that plants, which consume energy for both blooming and fruiting (female plants for dioecy, hermaphrodite for androdioecy), bear a higher cost of reproduction and hence ought to have a lower vegetative growth rate than males.

A number of comparisons have been made to examine the reproductive costs in sexual dimorphic species (Allen and Antos 1993; Cornelissen and Stiling 2005). However, less attention has been paid to the differences in biomass structure. If indeed resource allocation differs between sexes, then this difference might have some consequences, of which total biomass allocation pattern might be one (Zhang et al. 2012). Our results are consistent with those of previous studies. We found that the proportion of vegetative component biomass varies between genders, with males having a larger proportion of foliage and belowground biomass, implying that hermaphrodites have a different biomass allocation strategy than males (Fig. 2). It turns out that males allocate more resources to vegetative foliage and coarse roots.

4.2 Different biomass equations between genders

Biomass equations are an effective way for predicting biomass. The biomass of individual trees can be estimated from regression equations as a function of height, DBH, or other independent variables (Catchpole and Wheeler 1992; Zianis and Mencuccini 2003; Zheng et al. 2015). Ter-Mikaelian and Korzukhin (1997) presented 803 biomass equations for 65 species in North America. Bond-Lambery et al. (2002) reported that biomass allocation differs among components and established not only aboveground but also belowground biomass equations for six species in their temperate region. Biomass equations with DBH as an independent variable worked very well for the various components of A. tegmentosum and are consistent with previous studies for various other tree species (Ter-Mikaelian and Korzukhin 1997).

For androdioecious species, investigations into gender-independent biomass equations are rare. We opted for five types of equations to establish relations between biomass, vegetative and reproductive components, as well as tree attributes. Our results revealed different model forms for stem wood and total biomass (Tables 2 and 3) and different models for the same component biomass equation (except for branch biomass) between males and hermaphrodites (Table 4), suggesting that gender is a significant factor in biomass models. This result emphasized that, when establishing biomass equations for androdioecious species, sex roles are crucial in developing models.

4.3 Gender-related allometric relationships

As discussed earlier, males usually show a higher vegetative growth for vegetative components than hermaphrodites. Obeso (1998) found that males had a faster relative growth rate than females, which tended to have a lower relative growth rate in their fruit-bearing branches than males. Females have to allocate extra resources to successfully pollinated flowers. Similarly, Cornelissen and Stiling (2005) reported that male plants have better growth rates of vegetative organs. Our results show larger foliage and coarse root biomass in males than in hermaphrodites (Fig. 3 and Table 4), which agree with the findings of Popp and Reinartz (1988), Antos and Allen (1999), Rocheleau and Houle (2001), and Petzold et al. (2013).

For reproductive components, there were no significant differences in the number of flowers between genders, as shown in Fig. 4. However, the total reproductive investment in biomass was higher for hermaphrodites than for males, as a result of fruit production. Hermaphrodite A. tegmentosum plants have higher reproductive costs and lower vegetative biomass. Not surprisingly, this result accords with the predictions that there are clear tradeoffs between vegetative and reproductive matters.

Harris and Pannell (2010) reported that females of highly serotinous species were less branched than males in dioecious Leucadendron. Another study showed more branches and smaller leaves in males than in females (Midgley 2010). As the results in Figs. 3 and 4 demonstrate, hermaphrodites of A. tegmentosum have a lower value of foliage and coarse root biomass than males. In contrast, Zhang et al. (2012) reported that female plants have larger leaf, branch, and stem wood biomass than males, owing to their green flowers and fruits. In several studies, it has been demonstrated that green flowers and fruits are capable of photosynthesis and can supply a significant amount of their own carbon requirements (Bazzaz et al. 1979; Williams et al. 1985; Galen et al. 1993). For A. tegmentosum in this study, the reproductive components of hermaphrodite trees are also green, but they do not seem to contribute much to growth. García (1995) reported that the total amount of chlorophyll in fruits accounts for almost half the biomass of the aerial part of fruiting female plants, suggesting that photosynthetic activity of green fruits is far less than that of leaves. This phenomenon is likely to vary among species and depend on their reproductive features. Our results suggest that the sexual differences in allometric relationships might be related to differences in reproductive costs between hermaphrodites and males.